Tuesday, April 01, 2008

March Corbynwatch

A quick update on the March results: Corbyn gets a fail on the temperature (he predicted colder than normal, it was warmer according to here and here - note that "normal" in his terms is the 1961-90 mean, which was 5.67C) but a success on the rain - he said upwards of 130% of normal, and Philip Eden says we had 135% (I'm simplifying Corbyn's forecast here a bit, since I don't have precise regional data). As before, I'm ignoring the sunshine on the basis that this is essentially the converse of the rainfall forecast: for the record, he was actually wrong.

So a second successive 50% result means that the scores on the doors have "improved" to 2 out of 6 for the first 3 months of the year. However, despite the higher overall success rate, the failure of his method is actually marginally more clear-cut now, since the gap between the actual success rate and the claimed 80% forecast accuracy is more apparent:

Successes (out of 6)6 or 5 4 or 32 or less
Probability66%33%2%

(Yes there's a rounding error there, all three values are actually a little lower than quoted.)

I hope someone will supply April's forecast. If he gets the next 7 months spot-on (temp and precip) he'll be up to the 80% rate that he claims, and I will eat my hat, video the event and post it on Youtube.

3 comments:

P. Lewis said...

I guess what you're saying is that in another month, or maybe two, we can forget about PC entirely and your hat will be safe (not that it isn't already).

How would you do your hat anyway: sashimi or avec sushi? Or perhaps deep fried, like my northern Celtic bretheren like their Mars bars and pizza done (possibly together)?

EliRabett said...

Never thought anything could make me root for Corbyn. Well, being occasionally wrong is an occupational hazard.

James Annan said...

Definitely deep-fried pizza stylee, with lashings of salt 'n' soss.